Electric power information dynamic identification method based on deep learning

Abstract

The invention discloses an electric power information dynamic identification method based on deep learning, and relates to the technical field of electric power dynamic identification, and the method comprises regional power utilization monitoring, regional electric power scheduling, electric power scheduling monitoring and early warning prompt, and the method comprises the steps: monitoring the power utilization condition in a region, analyzing each scheduling region corresponding to a heavy pressure region, and carrying out the early warning prompt. And the stability of power transmission between each dispatching area and the heavy area is judged, so that automatic power utilization regulation and control in the area are realized, reasonable application of power in the area is guaranteed, the power in the area is in a balanced state, the problem of power shortage in the area is solved, and the safety and stability of power utilization in the area are further guaranteed. And meanwhile, on the basis of feature monitoring and extraction of the power information in the region based on the deep learning technology, the recognition speed of the power state in the region can be effectively improved, a reasonable power borrowing distribution scheme is formulated in the region, and safe and efficient application of power in the region is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of dynamic power identification, and in particular to a method for dynamic power information identification based on deep learning. Background Art

[0002] In today's life, the demand for electricity is getting higher and higher, so it is necessary to monitor the electricity consumption, and then realize intelligent and automatic electricity consumption regulation to ensure the rational application of electricity. By monitoring the electricity consumption in each area, the negative pressure situation of electricity application in each area can be obtained, so as to implement regional electricity consumption regulation and solve the problem of electricity shortage in the region. At the same time, the power dispatching situation corresponding to each dispatching area and the heavy pressure area is monitored to ensure the stability of electricity transfer between the heavy pressure area and each transmission area, formulate a reasonable transfer allocation for electricity, alleviate the electricity shortage in the region, and make intelligent application arrangements for electricity, thereby improving the availability of electricity and supporting application scope.

[0003] Prior art, such as the invention patent application with announcement number CN107240033B, discloses a method and system for constructing an electric power identification model. This method determines the customer characteristics corresponding to each predetermined electric power customer type (at least including the potential complaint-prone customer type) based on the electric power historical business data, and determines the various model factors required to construct the electric power identification model. By analyzing the influence of each model factor on the customer type identification result, and based on the influence analysis result, the comprehensive weight of each model factor is obtained, and finally an electric power identification model including each model factor and the model factor weight is obtained. This model can be used to identify the type of electric power customers, and then identify the potential complaint-prone customers in the electric power industry. It can be seen that the application of this application scheme can realize the use of the constructed electric power identification model to more intuitively, accurately and quickly identify the potential complaint-prone customers in the electric power industry, thereby effectively solving the various problems existing in the way of recognizing customers through subjective judgment.

[0004] In view of the above scheme, the applicant of the present invention found that the above technical problems have at least the following technical problems: the above invention mainly uses the constructed power identification model to more intuitively, accurately and quickly identify potential complaint-prone customers in the power industry, thereby effectively solving various problems existing in the way of recognizing customers through subjective judgment. It is not based on the power consumption in each area, and then understands the negative pressure state of the applied power in each area. It is impossible to perform corresponding automatic power consumption regulation according to the applied power situation in each area, and it is impossible to achieve intelligent and reasonable power supply guarantee for the current area. It is impossible to quickly know the heavy-pressure areas in each area, and it is impossible to timely perform timely power consumption regulation for the heavy-pressure areas, and increase the power consumption pressure in the heavy-pressure areas. At the same time, it does not monitor and analyze the transmission of power dispatching between each dispatching area and the heavy-pressure area, and it is impossible to ensure whether the power dispatching between each dispatching area and the heavy-pressure area is in a stable transmission state, and it is impossible to achieve power dispatching transmission with the highest adaptability, and it is impossible to ensure intelligent power supply and demand in the area, increase the power consumption pressure in the area, and lead to power imbalance in the area. Summary of the invention

[0005] In view of the above-mentioned technical deficiencies, the purpose of the present invention is to provide a method for dynamic identification of power information based on deep learning.

[0006] In order to solve the above technical problems, the present invention adopts the following technical solutions: The present invention provides a method for dynamic identification of power information based on deep learning, including: Step 1, regional power consumption monitoring: By monitoring the power consumption in the region, and then analyzing the negative pressure state of power consumption in each region.

[0007] Step 2: Regional power dispatch: Obtain the electricity data corresponding to each region, and then analyze and obtain the stability results of the corresponding power values ​​of each region. At the same time, monitor the sub-regions with qualified stability results corresponding to the power values, analyze and obtain the dispatching areas corresponding to the heavy-pressure areas, and analyze and obtain the dispatching area supply corresponding to the power dispatch of the heavy-pressure areas.

[0008] Step 3: Power dispatch monitoring: monitor the power transmission data between each dispatch area and the heavy pressure area, and then determine the stability of power transmission between each dispatch area and the heavy pressure area.

[0009] Step 4: Early warning: When a certain area has heavy negative pressure on the applied electricity, an early warning will be issued.

[0010] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention provides a method for dynamic identification of power information based on deep learning, which monitors the power consumption in the region, analyzes the dispatching areas corresponding to the heavy-pressure area, and judges the stability of power transmission between the dispatching areas and the heavy-pressure area, so as to realize automatic power consumption regulation in the region, ensure the rational application of power in the region, make the power in the region in a balanced state, solve the problem of power shortage in the region, further ensure the safety and stability of power consumption in the region, solve the shortcomings in the current technology, and at the same time, based on the deep learning technology, the feature monitoring and extraction of power information in the region can effectively improve the recognition speed of the power status in the region, reduce the pressure of manual power monitoring, and avoid the misjudgment rate of power status identification in the region due to human factors. The automation function based on deep learning can provide a quick solution for the uneven power in the region, formulate a reasonable power transfer allocation for the region, alleviate the power consumption pressure in the area with large power demand, realize intelligent regional power application arrangement, improve the availability of power and support application scope, and ensure the safe and efficient application of power in the region. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0012] Figure 1 The present invention is a schematic flow chart of the steps for implementing the method. DETAILED DESCRIPTION

[0013] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0014] See also Figure 1 As shown, a method for dynamic identification of power information based on deep learning includes: Step 1, regional power consumption monitoring: by monitoring the power consumption in the region, and then analyzing to obtain the negative pressure state of power consumption in each region.

[0015] It should be noted that the historical division distance corresponding to the area is obtained from the database, and the average value of the historical division distance is calculated to obtain the average adjacent distance, and the average adjacent distance is used as the reference standard for the preset reference adjacent distance. The reference adjacent distance is preset in this way, and the corresponding historical electricity consumption value in the area is obtained from the database, and the average value of the corresponding historical electricity consumption value in the area is calculated to obtain the corresponding historical average electricity consumption value in the area, and the corresponding historical average electricity consumption value in the area is used as the historical reference electricity consumption value.

[0016] As an optional implementation, the analysis obtains the negative pressure state of each area corresponding to the applied electricity, and the specific analysis process is as follows: B1. Obtain the corresponding area image in the area through the camera, and extract the corresponding adjacent distances between each main device from the image, and preset a reference adjacent distance, and compare the corresponding adjacent distances between each main device with the preset adjacent distance. If the adjacent distance between a main device and a main device is less than or equal to the preset reference adjacent distance, then the main device and the main device are divided into one area, and so on to obtain each area.

[0017] B2. Obtain the power consumption corresponding to each area from the smart meters corresponding to each area, and compare it with the historical reference power consumption values ​​stored in the database. If the power consumption corresponding to a certain area is less than or equal to the historical reference power consumption stored in the database, it is determined that the negative pressure of the power application in this area is small. Conversely, if the power consumption corresponding to a certain area is greater than the historical reference power consumption stored in the database, it is determined that the negative pressure of the power application in this area is large, and the area is recorded as a heavy pressure area. The negative pressure state of each area for power application is obtained by analysis.

[0018] Step 2: Regional power dispatch: Obtain the electricity data corresponding to each region, and then analyze and obtain the stability results of the corresponding power values ​​of each region. At the same time, monitor the sub-regions with qualified stability results corresponding to the power values, analyze and obtain the dispatching areas corresponding to the heavy-pressure areas, and analyze and obtain the dispatching area supply corresponding to the power dispatch of the heavy-pressure areas.

[0019] It should be noted that the corresponding model algorithms in deep learning include feedforward neural networks, convolutional neural networks, recurrent neural networks, and generative adversarial networks.

[0020] It should be noted that the sensors include digital harmonic analyzers, smart meters and load monitors, etc.

[0021] As an optional implementation method, the specific analysis process of obtaining the electricity data corresponding to each region is as follows: by obtaining the electricity data of each region when applying electricity, and obtaining from the database a historical area that is under tight electricity pressure and needs to allocate electricity, the reference evaluation data of the electricity stability of each historical area is used, and the electricity data features corresponding to each electricity data in each region are extracted based on the reference evaluation data, and the extracted electricity data features are processed normally, and the corresponding model algorithm in deep learning is selected, and the electricity data features are automatically generated based on the selected model algorithm to obtain the corresponding electricity feature recognition model, and the electricity feature recognition model is subjected to corresponding model training, and at the same time, each sensor is used to obtain in real time the electricity data corresponding to the current area when executing electricity supply, wherein the electricity feature data includes frequency fluctuation, number of power outages and load fluctuation.

[0022] As an optional implementation, the analysis obtains stability results of the electric quantity values ​​corresponding to each region, and the specific analysis process is as follows: the electric data corresponding to each region is input into the electric power feature recognition model, and then the stability results of the electric quantity values ​​corresponding to each region are output. If the stability output result of the electric quantity values ​​corresponding to a region is 0, it is determined that the electric quantity distribution application corresponding to the region is in a stable state, and the region is recorded as a sub-region. If the stability output result of the electric quantity values ​​corresponding to a region is 1, it is determined that the electric quantity distribution application corresponding to the region is tense and in an unstable state, and it is determined that the region is not suitable for power adjustment to the heavy pressure region, so as to obtain the stability results of the electric quantity values ​​corresponding to each region through analysis.

[0023] As an optional implementation, the output obtains the stability result of the power value corresponding to each area, and the specific output process is as follows: by calculating the formula: , calculate the first Stability results of the corresponding power values ​​in each area , is the number of each area, , is any integer greater than 2, For the The frequency fluctuation of the power value in each area. For the The number of power outages in each area corresponding to the power value, For the The load fluctuation of the corresponding power value in each area, , , They are the weight factors of the set frequency fluctuation, the power outage number, and the load fluctuation. For reference stability power value, , , .

[0024] It should be noted that the historical stability power parameters of the corresponding power values ​​of each region in a stable state are obtained from the database, and the historical stability power parameters of the corresponding power values ​​of each region in a stable state are statistically analyzed, and the average value of the statistically analyzed historical stability power parameters is calculated to obtain the historical average stability power parameters, which are used as reference stability power values.

[0025] It should also be noted that the principal component analysis method is used to obtain the weight factors of frequency fluctuations, the weight factors of the number of power outages, and the weight factors of load fluctuations. The corresponding information of frequency fluctuations, power outages, and load fluctuations is first condensed, and then the corresponding variance explanation rate is obtained, and then divided by the variance explanation rate to obtain the corresponding weight.

[0026] It is important to note again that information condensation refers to condensing multiple analysis items into several key summary indicators, and the variance explanation rate refers to the contribution of each principal component or factor to the overall variance, usually expressed as a percentage, indicating how much information the factor contains about the original data.

[0027] As an optional implementation method, the analysis obtains the scheduling areas corresponding to the heavy pressure areas, and the specific analysis process is as follows: A1. Arrange the stability results in order from high to low, and then obtain the sub-areas arranged in order, and at the same time obtain the remaining power corresponding to each sub-area from the distribution meter corresponding to each sub-area.

[0028] A2. By calculation formula: , and then calculate the The power supply value corresponding to each sub-area , is the number of each sub-region, , is any integer greater than 2, is the reference remaining power set. For the The remaining power corresponding to each sub-area is For the The stability results of the power values ​​corresponding to each area.

[0029] It should be noted again that the historical remaining power corresponding to each sub-area is obtained from the database, and the historical remaining power corresponding to each sub-area is calculated based on the average value to obtain the average historical remaining power corresponding to each sub-area, and the average historical remaining power corresponding to each sub-area is used as the reference remaining power.

[0030] A3. Compare the available power value corresponding to each sub-area with the reference available power value stored in the database. If the available power value corresponding to a sub-area is greater than or equal to the reference available power value stored in the database, it is determined that the sub-area can execute the corresponding power dispatch, and the sub-area is recorded as the dispatch area. On the contrary, if the available power value corresponding to a sub-area is less than the reference available power value stored in the database, it is determined that the sub-area cannot execute the corresponding power dispatch, and so on, the dispatch areas corresponding to the heavy pressure area are analyzed and obtained.

[0031] As an optional implementation method, the analysis obtains the supply of the dispatching area corresponding to the power dispatch of the heavy-stress area, and the specific analysis process is as follows: based on the remaining power corresponding to each dispatching area and the basic power reserved for its own power supply operation, the basic power is subtracted from the remaining power to obtain the dispatchable power, and so on to obtain the dispatchable power corresponding to each dispatching area, and based on the stability results corresponding to each dispatching area, the dispatchable values ​​corresponding to each dispatching area are arranged in order from high to low, and then the demand power corresponding to the heavy-stress area is obtained. When the sum of the dispatchable power between a certain dispatching area and the dispatching area is greater than the demand power corresponding to the heavy-stress area, the dispatching area corresponding to the sum is used as the first dispatching area and the second dispatching area of ​​the heavy-stress area, and so on to obtain the sequential dispatching areas corresponding to the power dispatch of the heavy-stress area, so as to obtain the dispatching area supply of the power dispatch corresponding to the heavy-stress area.

[0032] Step 3: Power dispatch monitoring: monitor the power transmission data between each dispatch area and the heavy pressure area, and then determine the stability of power transmission between each dispatch area and the heavy pressure area.

[0033] It should be noted that the sensors include power quality analyzers, oscilloscopes and electric energy meters, etc.

[0034] As an optional implementation method, the monitoring of the power transmission data between each dispatching area and the heavy pressure area, the specific monitoring process is as follows: by obtaining various types of power transmission data between each dispatching area and the heavy pressure area, and obtaining the reference data for judging the transmission stability during the historical power transmission between each dispatching area and the heavy pressure area stored in the database, based on the reference data, extracting the corresponding transmission data features from the various types of power transmission data between the current dispatching areas and the heavy pressure area, and performing corresponding standardization processing on the extracted transmission data features, and selecting the corresponding model algorithm for deep learning, inputting the transmission data features into the model algorithm, constructing the corresponding dispatching power supervision model, and performing corresponding model prediction training on the power supervision model, and at the same time using sensors to obtain in real time the power transmission data of each current dispatching area and the heavy pressure area during power dispatch, wherein the power transmission data includes line loss, transient voltage and hourly load.

[0035] As an optional implementation manner, the stability of the power transmission between each dispatching area and the heavy pressure area is judged, and the specific judgment process is as follows: the power transmission data between each dispatching area and the heavy pressure area is input into the dispatching power supervision model, and then the stability coefficient of the power transmission between each dispatching area and the heavy pressure area is output. If the output value of the stability coefficient of the power transmission between a certain dispatching area and the heavy pressure area is 0, it is judged that the power transmission between the dispatching area and the heavy pressure area is in a stable state, and the corresponding power dispatching transmission is maintained. If the output value of the stability coefficient of the power transmission between a certain dispatching area and the heavy pressure area is 1, it is judged that the power transmission between the dispatching area and the heavy pressure area is in an unstable state, and the dispatching transmission of power between the dispatching area and the heavy pressure area is stopped, and based on the dispatching area supply corresponding to the power dispatch of the heavy pressure area, the power dispatching transmission of the dispatching area behind the unstable dispatching area is started, so as to judge the stability of the power transmission between each dispatching area and the heavy pressure area.

[0036] As an optional implementation, the output obtains the stability coefficient of the power transmission between the dispatching area and the heavy pressure area. The specific output process is as follows: by calculating the formula: , the output is Stability coefficient of power transmission between the dispatching area and the heavy pressure area , is the number of each dispatch area, , is any integer greater than 2, For the The line loss of power transmission between the dispatching area and the heavy pressure area, For the The transient voltage of power transmission between the dispatching area and the heavy pressure area, For the The hourly load of electricity transmission between the dispatching area and the heavy pressure area, , , They are the weight factors of the set line loss, transient voltage and hourly load, respectively. , , , is the set reference transmission stability coefficient.

[0037] It should be noted that the historical transmission stability coefficient of the power transmission between each dispatching area and the heavy pressure area in a stable state is obtained from the database, and the historical transmission stability coefficient of the power transmission between each dispatching area and the heavy pressure area in a stable state is statistically analyzed, and the historical transmission stability coefficient is averaged to obtain the historical average transmission stability coefficient, which is used as the reference transmission stability coefficient.

[0038] It should also be noted that the weight factors of line loss, transient voltage and hourly load are obtained through principal component analysis. The information of corresponding line loss, transient voltage and hourly load is condensed first, and then the corresponding variance explanation rate is obtained, and then divided by the variance explanation rate to get the corresponding weight.

[0039] Step 4: Early warning: When a certain area has heavy negative pressure on the applied electricity, an early warning will be issued.

[0040] The embodiment of the present invention monitors the power consumption in the area, analyzes the dispatching areas corresponding to the heavy-stress area, and determines the stability of power transmission between the dispatching areas and the heavy-stress area, so as to realize automatic power consumption regulation in the area, ensure the rational application of power in the area, make the power in the area in a balanced state, solve the problem of power shortage in the area, further ensure the safety and stability of power consumption in the area, and solve the deficiencies in the current technology. At the same time, based on the deep learning technology, the feature monitoring and extraction of power information in the area can effectively improve the recognition speed of the power status in the area, reduce the pressure of manual power monitoring, and avoid the misjudgment rate of power status recognition in the area due to human factors. The automation function based on deep learning can provide a quick solution for the uneven power in the area, formulate a reasonable power transfer allocation for the area, alleviate the power consumption pressure in the area with large power demand, realize intelligent regional power application arrangement, improve the availability of power and support application scope, and ensure the safe and efficient application of power in the area.

[0041] The above contents are merely examples and explanations of the concept of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined in this specification, they shall all fall within the protection scope of the present invention.

Claims

1. A method for dynamic identification of power information based on deep learning, characterized in that: include: Step 1: Regional power consumption monitoring: By monitoring the power consumption in the region, the negative pressure state of power consumption in each region can be analyzed; Step 2: Regional power dispatch: Obtain the power data corresponding to each region, and then analyze and obtain the stability results of the power values ​​corresponding to each region. At the same time, monitor the sub-regions with qualified stability results corresponding to the power values, analyze and obtain the dispatching areas corresponding to the heavy-pressure areas, and analyze and obtain the dispatching area supply corresponding to the power dispatch of the heavy-pressure areas; Step 3: Power dispatch monitoring: monitor the power transmission data between each dispatch area and the heavy pressure area, and then determine the stability of power transmission between each dispatch area and the heavy pressure area; Step 4: Early warning: When a certain area has heavy negative pressure on the applied electricity, an early warning will be issued.

2. A method for dynamic identification of power information based on deep learning as claimed in claim 1, characterized in that: The analysis obtains the negative pressure state of each area corresponding to the applied electricity. The specific analysis process is as follows: B1. Obtain the corresponding regional image in the area through the camera, extract the corresponding adjacent distances between the main devices from the image, preset the reference adjacent distance, compare the corresponding adjacent distances between the main devices with the preset adjacent distance, and if the adjacent distance between a main device and a main device is less than or equal to the preset reference adjacent distance, divide the main device and the main device into one area, and divide each area by analogy; B2. Obtain the power consumption corresponding to each area from the smart meters corresponding to each area, and compare it with the historical reference power consumption values ​​stored in the database. If the power consumption corresponding to a certain area is less than or equal to the historical reference power consumption stored in the database, it is determined that the negative pressure of the power application in this area is small. Conversely, if the power consumption corresponding to a certain area is greater than the historical reference power consumption stored in the database, it is determined that the negative pressure of the power application in this area is large, and the area is recorded as a heavy pressure area. The negative pressure state of each area for power application is obtained by analysis.

3. A method for dynamic identification of power information based on deep learning as claimed in claim 1, characterized in that: The specific analysis process of obtaining the electrical data corresponding to each area is as follows: By obtaining the power data of each region when applying electricity, and obtaining from the database the historical data of a region that is under tight electricity pressure and needs to allocate electricity, the reference evaluation data of the power stability of each region in history is used, and the power data features corresponding to each power data in each region are extracted based on the reference evaluation data, and the extracted power data features are processed by normalization, and the corresponding model algorithm in deep learning is selected. The power data features are automatically generated based on the selected model algorithm to obtain the corresponding power feature recognition model, and the power feature recognition model is trained accordingly. At the same time, various sensors are used to obtain the corresponding electricity data of each region when executing power supply in real time, where the power feature data includes frequency fluctuations, number of power outages and load fluctuations.

4. The method for dynamic identification of power information based on deep learning according to claim 1, characterized in that: The analysis obtained the stability results of the corresponding power values ​​in each area. The specific analysis process is as follows: The electricity data corresponding to each area is input into the power feature recognition model, and then the stability results of the corresponding electricity values ​​of each area are output. If the stability output result of the corresponding electricity value of a certain area is 0, it is determined that the electricity distribution application corresponding to the area is in a stable state, and the area is recorded as a sub-area. If the stability output result of the corresponding electricity value of a certain area is 1, it is determined that the corresponding electricity distribution application of the area is tense and in an unstable state, and it is determined that the area is not suitable for power adjustment to the heavy-pressure area. The stability results of the corresponding electricity values ​​of each area are obtained by analysis.

5. A method for dynamic identification of power information based on deep learning as claimed in claim 4, characterized in that: The output obtains the stability result of the corresponding power value of each area, and the specific output process is as follows: By calculation formula: , calculate the first Stability results of the corresponding power values ​​in each area , is the number of each area, , is any integer greater than 2, For the The frequency fluctuation of the power value in each area. For the The number of power outages in each area corresponding to the power value, For the The load fluctuation of the corresponding power value in each area, , , They are the weight factors of the set frequency fluctuation, the power outage number, and the load fluctuation. For reference stability power value, , , .

6. A method for dynamic identification of power information based on deep learning as claimed in claim 1, characterized in that: The analysis obtains the dispatching areas corresponding to the heavy pressure areas. The specific analysis process is as follows: A1. Arrange the stability results in descending order to obtain the sub-areas arranged in order, and obtain the remaining power corresponding to each sub-area from the power distribution meter corresponding to each sub-area; A2. By calculation formula: , and then calculate the The power supply value corresponding to each sub-area , is the number of each sub-region, , is any integer greater than 2, is the reference remaining power set. For the The remaining power corresponding to each sub-area is For the The stability results of the power values ​​corresponding to each area; A3. Compare the available power value corresponding to each sub-area with the reference available power value stored in the database. If the available power value corresponding to a sub-area is greater than or equal to the reference available power value stored in the database, it is determined that the sub-area can execute the corresponding power dispatch, and the sub-area is recorded as the dispatch area. On the contrary, if the available power value corresponding to a sub-area is less than the reference available power value stored in the database, it is determined that the sub-area cannot execute the corresponding power dispatch, and so on, the dispatch areas corresponding to the heavy pressure area are analyzed and obtained.

7. The method for dynamic identification of power information based on deep learning according to claim 1, characterized in that: The analysis obtains the dispatching area supply corresponding to the heavy pressure area. The specific analysis process is as follows: Based on the remaining power corresponding to each dispatching area and the basic power required to retain its own power supply operation, the basic power is subtracted from the remaining power to obtain the dispatchable power. Similarly, the dispatchable power corresponding to each dispatching area is obtained. Based on the stability results corresponding to each dispatching area, the dispatchable values ​​corresponding to each dispatching area are arranged in order from high to low, and then the demand power corresponding to the heavy-pressure area is obtained. When the sum of the dispatchable power between a certain dispatching area and the dispatching area is greater than the demand power corresponding to the heavy-pressure area, the dispatching area corresponding to the sum is used as the first dispatching area and the second dispatching area of ​​the heavy-pressure area. By analogy, the sequential dispatching areas corresponding to the power dispatch of the heavy-pressure area are obtained, and the dispatching area supply corresponding to the power dispatch of the heavy-pressure area is obtained by analysis.

8. The method for dynamic identification of power information based on deep learning according to claim 1, characterized in that: The specific monitoring process of monitoring the power transmission data between each dispatching area and the heavy pressure area is as follows: By obtaining various types of power transmission data between each dispatching area and the heavy-pressure area, and obtaining the reference data for judging the transmission stability during the historical power transmission between each dispatching area and the heavy-pressure area stored in the database, the corresponding transmission data features are extracted from the various types of power transmission data between the current dispatching areas and the heavy-pressure areas based on the reference data, and the extracted transmission data features are standardized accordingly, and the corresponding model algorithm for deep learning is selected, and the transmission data features are input into the model algorithm to construct the corresponding dispatching power supervision model, and the power supervision model performs the corresponding model prediction training, and at the same time, the sensors are used to obtain the power transmission data of each current dispatching area and the heavy-pressure area during power dispatch in real time, where the power transmission data includes line loss, transient voltage and hourly load.

9. A method for dynamic identification of power information based on deep learning as claimed in claim 8, characterized in that: The specific judgment process of judging the stability of power transmission between each dispatching area and the heavy pressure area is as follows: The power transmission data between each dispatching area and the heavy pressure area is input into the dispatching power supervision model, and then the stability coefficient of the power transmission between each dispatching area and the heavy pressure area is output. If the output value of the stability coefficient of the power transmission between a dispatching area and the heavy pressure area is 0, it is determined that the power transmission between the dispatching area and the heavy pressure area is in a stable state, and the corresponding power dispatching transmission is maintained. If the output value of the stability coefficient of the power transmission between a dispatching area and the heavy pressure area is 1, it is determined that the power transmission between the dispatching area and the heavy pressure area is in an unstable state, and the dispatching transmission of power between the dispatching area and the heavy pressure area is stopped, and based on the dispatching area supply corresponding to the power dispatch of the heavy pressure area, the power dispatching transmission of the dispatching area behind the unstable dispatching area is started, so as to judge the stability of the power transmission between each dispatching area and the heavy pressure area.

10. A method for dynamic identification of power information based on deep learning as claimed in claim 9, characterized in that: The output obtains the stability coefficient of power transmission between the dispatching area and the heavy pressure area. The specific output process is as follows: , the output is Stability coefficient of power transmission between the dispatching area and the heavy pressure area , is the number of each dispatch area, , is any integer greater than 2, For the The line loss of power transmission between the dispatching area and the heavy pressure area, For the The transient voltage of power transmission between the dispatching area and the heavy pressure area, For the The hourly load of electricity transmission between the dispatching area and the heavy pressure area, , , They are the weight factors of the set line loss, transient voltage and hourly load, respectively. , , , is the set reference transmission stability coefficient.

Images